Supervisor: Torsten Wik, Department of electrical engineering Examiner: Britt-Marie Wilén, Department of Water Environment Technology
Master’s Thesis ACEX30-18-91 Department of Arcitechure and Civil Engineering Division of Water Environment Technology Chalmers University of Technology SE-412 96 Gothenburg Telephone +46 31 772 1000
Cover: Clarias Gariepinus Typeset in LATEX Gothenburg, Sweden 2018 iv
Analysis of a recirculating aquaculture system An analysis at Lantfisk Master’s Thesis in the Master’s programme Innovative and Sustainable Chemical Engineering AMANDA ANDERSSON MÅNS GERDTSSON Department of Civil and Environmental Engineering Division of Water Environment Technology Chalmers University of Technology
The water treatment in a commercial RAS used for production of Clarias Gariepinus was studied in order to gain understanding of the efficiency of the process. In order to evaluate the capacity of the water treatment several methods were used such as; analysis of nitrogen compounds with ion chromatography, analysis of total organic carbon, microscopic investigation of sludge, analysis of COD and BOD and activity tests of nitrifying and denitrifying bacteria. It was found that the concentration difference of the nitrogen compounds between the incoming and outgoing flow of the treatment process were small due to low activity and short retention times. No concentrations of the nitrogen compounds exceeded the limit values for what the fish can withstand. However, the water has high COD and very low BOD. Carbon should be removed in order to improve nitrification while the denitrification is limited by the low amount of biodegradable carbon. It was also found that the sludge in the pump sumps performed better in the activity test than the sludge from the denitrification tanks. Although the water treatment process of the RAS has some areas of improvements, the process has shown to be insensitive to disruptions and able to recover from interference.
Keywords: Recirculating aquaculture system, RAS, Clarias Gariepinus, nitrification, denitrification. v
Acknowledgements Thanks to our examiner Britt-Marie Wilén and our supervisor Torsten Wik for their support throughout the project. Thanks to Diana Olsson Waage at Lantfisk for letting us use their facility for the purpose of this study. Also thanks to Robin Ek and Kalle Larsson for their assistance during the work at lantfisk. Special thanks to Mona Pålsson for her assistance during laboratory work at the Environmental Chemistry Laboratory at Chalmers university of technology.
Amanda Andersson and Måns Gerdtsson, Gothenburg, June 2018
Contents List of Figures
List of Tables
1 Introduction 1 1.1 Fish production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1.1 Recirculating aquaculture systems RAS . . . . . . . . . . . . . 1 1.2 Lantfisk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.2.1 RAS at lantfisk . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.3 Research questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.3.1 What is the nitrogen removal rate? . . . . . . . . . . . . . . . 4 1.3.2 Are there daily variations of nitrogen compounds in the system? 5 1.3.3 What is the amount of dissolved carbon in the system and how much of it is biodegradable? . . . . . . . . . . . . . . . . 5 1.3.4 Is it viable to operate a RAS without a dedicated sludge removal unit? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2 Theory 2.1 Water treatment in RAS 2.2 Ion chromatography . . 2.3 Carbon removal . . . . . 2.4 Flow . . . . . . . . . . . 2.5 Excretion . . . . . . . .
Dimensions of tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . Components of studied RAS at Lantfisk . . . . . . . . . . . . . . . . Average retention times. Since there are three parallel lines for DN and OCR the total retention time is shown for a single line. Fish tanks are also connected in parallel and retention time is given as the average for a single tank . . . . . . . . . . . . . . . . . . . . . . . . . Nitrogen excretion based on feed rate during the series. The concentration increase is based on the volume of the entire system. . . . . Rate of nitrification, NF1: First nitrification tank, NF5: last nitrification tank, L14: First OCR tank. . . . . . . . . . . . . . . . . . . Rate of denitrification. . . . . . . . . . . . . . . . . . . . . . . . . . Concentrations of metals (mg/kg dry matter) . . . . . . . . . . . . Weights of dry sludge samples . . . . . . . . . . . . . . . . . . . . .
. 17 . . . .
23 27 35 35
List of Tables
1 Introduction 1.1
The fish production, both aquaculture and capture production, has grown significantly since the 1950’s and must grow further to satisfy the increasing global population and consumption. In 2013, the total fish production reached a number of 162.9 million tonnes of which 141.5 million tonnes was used for human consumption. The estimated global annual fish consumption per capita has increased from 9.9 kg in the 1960s to 14.4 kg in the 1990s to 19.7 kg in 2013. This is partly because of increasing production of fish but also due to better distribution to consumers and better utilization of the product, which reduces waste. This fast increase of fish production demands sustainable strategies and techniques for fishing that take social, economical and environmental aspects into consideration. At some places around the world, the capture fisheries production has reached a point where it risks extinction of local fish stocks. This could in turn lead to disruption of ecosystems and devastate the subsistence for people who depend on fishing. However, improvements of the of the fisheries management has led to small amends in the state of some fish stocks. The increased growth in aquaculture stands for almost half of the human consumption of fish. The most common method for traditional fish farming is in open cage systems in the ocean or in lakes. Large fish cages are placed in already existent lakes or in the ocean where they utilize the surrounding ecosystem for water flow into the system and transport of faeces and food waste out of the system. This is a cost effective and well established method but it also causes strain on the ecosystem because of the nutrients and particles spread to the local environment. There is also a risk of spreading disease and escape of fish, which could disrupt the already existing ecosystem.
Recirculating aquaculture systems RAS
Semi-closed or closed systems have been developed to reduce the environmental impact of the open cage systems. This technique for fish farming can also be placed in lakes or in the ocean. Water is pumped into a closed container with fish, which can be a “moving bag” or a solid tank, and the water flows out from the container at specific outlets. The water is then processed in a water treatment plant and can be returned to surrounding water or a closed container for the fish.
One method of fish farming that gives better control of the water treatment process is RAS, recirculated aquaculture system. It is a land based process that implements biological water treatment processes that removes nitrogen, biological matter and phosphorus. This enables a high degree of water to be recirculated and reused in the fish tanks. Nitrogen removal is important since the fish excrete ammonia from their gills and ammonia is toxic to the fish at high concentrations. Nitrogen removal is achieved by the processes called nitrification and denitrification, which is further explained in section 2.1. The sludge produced in the process can be removed and sent to sewage treatment or used as fertilizer. Compared to open cage system, RAS has many advantages, such as reduction of pathogenic bacteria and disease, low water use and high control of operational parameters. It also enables fish farming in areas where the access to water is poor. On the other hand, it is an expensive process, both in investment cost and in operational cost. It also requires close control by experienced staff since the system is sensitive to changes in process parameters.
Lantfisk is a small but expanding company on the outskirts of Gothenburg that utilizes the RAS technique to farm Clarias gariepinus, also calles African sharptooth catfish. They started their business in 2013 at a very small scale and in 2017 they produced 24 tonnes of fish. In 2018 they are planning on expanding their production even further and expect to produce 40 tonnes of fish. Since Lantfisk aims at continuous expansion of their production they wish to gain further knowledge about their RAS.
RAS at lantfisk
The flow chart of the RAS at Lantfisk is shown in Figure 1.1a. Floating feed is provided with automatic feeders from 06:00 to 17:00. The tanks labeled NF and OCR are aerated with pressurized air, which also cause agitation. The process is a closed system, which means that all the water is recirculating within the system. It is only refilled with water that corresponds to the loss of evaporation. More loops than expected were found in the system. The loops have been introduced in order to increase operating safety. Mainly the risk of overflow has decreased according to Lantfisk. As is shown in Figure 1.1a water leaving Pump 2 can either pass through the anoxic denitrification tanks (DN) and the aerated organic carbon removal tanks (OCR), or pass directly to the OCR tanks. This bypass is introduced in order to avoid overflow in the DN tanks while maintaining a high flow through the OCR tanks in order to aerate the water to provide sufficient oxygen to the fish.
(b) Flow through pumps (a) RAS flow chart. DN=anaerobic tanks for deni-
trification, OCR=aerated tanks for organic carbon removal, NF=aerated tanks for nitrification. The number of tanks in series is also indicated for each unit.
Pump 1 Pump 2 Pump 3
Flow (l/s) 2.0 2.1 8.2
The bioreactors used for the water treatment are filled with Kaldnes bio carriers in order to provide sufficient area for microorganisms to grow. In the tanks labeled NF and OCR in Figure 1.1a the bio carriers are moving around in the water as a result of the aeration. The bio carriers in the tanks labeled DN are stationary because the these tanks are filled with more carriers than the others, the flow is lower and there is no aeration. This effectively turns these tanks into fixed bed bio reactors. The water treatment of RAS is discussed further in section 2.1 Since Pump 3 has a higher flow than Pump 2 most of the water from the fish is recirculated back through the pump sumps and does not reach the treatment. All the tanks, including the pump sumps has the same dimensions, see Table 1.1. In Table 1.2 the components of the system is listed. Table 1.1: Dimensions of tanks Height (m) 1
Length (m) 1.2
Width (m) Volume (m3 ) 1 1.2
1. Introduction Table 1.2: Components of studied RAS at Lantfisk
Number of units (n) Average water level in units (m) Total volume in units (m3 ) Ratio of component to entire system (%)
Retention times in different tanks are calculated according to Equation 2.4. The retention time vary between the units and is shown as averages in Table 1.3. Since the denitrifying tanks are not agitated the hydraulic retention time is not a good approximation of the residence time. However, the flow rate is 3-4 times lower into the denitrifying tanks than into the OCR tanks. Table 1.3: Average retention times. Since there are three parallel lines for DN and OCR the total retention time is shown for a single line. Fish tanks are also connected in parallel and retention time is given as the average for a single tank Individual tanks (min) Total (min) Nitrification 5,9 30,0 Organic carbon removal 13,2 26,4 Fish tanks 33,5 There is no dedicated unit for removal of solids and most of the solids are trapped in the denitrification tanks where the flow is lowest and there is no agitation. These tanks fill up with solids and are therefore emptied approximately once a month. Solids also settle in the pump sumps. This creates anoxic environments where denitrification can occur both in the denitrification tanks and in the pump sumps.
The following are research questions that this project was aiming to answer.
What is the nitrogen removal rate?
The fish excrete ammonium which is toxic and has to be removed in a recirculating system. The removal rates of ammonium and nitrate have therefore been studied.
Are there daily variations of nitrogen compounds in the system?
The fish is only fed during parts of the day. This could for example result in lower concentrations of waste in the morning than at night.
What is the amount of dissolved carbon in the system and how much of it is biodegradable?
The amount of dissolved carbon in the water was expected to be high in the entire system because the water has a brown colour. The majority of the dissolved carbon is also expected to not be digestible by the microorganisms. An aim has therefore been to determine the amount of carbon in the system, and if it is biodegradable.
Is it viable to operate a RAS without a dedicated sludge removal unit?
The system has no dedicated sludge removal unit. Instead sludge builds up in the denitrification tanks where the flow is low and there is no agitation. When there is too much sludge in the denitrification tanks they are emptied and are therefore used for both sludge removal and denitrification.
2 Theory 2.1
Water treatment in RAS
An efficient water treatment process is crucial for RAS. Ammonium should be kept at a level below 45 mgN H4 − N/l and nitrate below 140 mgN O3 − N/l in order to avoid disturbances in physiology, growth and feed intake .There are several different RAS setups for fish production and the one that Lantfisk based their system on is shown in Figure 2.1.
Figure 2.1: Theoretical RAS setup. The conventional RAS configuration uses nitrifying biofilters to reduce ammonia and nitrite concentrations by oxidizing them into nitrate. This is combined with organic carbon removal where organic matter remaining after denitrification is removed by heterotrophic bacteria in aerobic tanks. The sludge created in this process can be removed by sedimentation or mechanical filtration . The nitrification is carried out by ammonia oxidizing bacteria (AOB) and nitrite oxidizing bacteria (NOB) in aerobic tanks according to the following reactions 2.1 and 2.2. These bacteria are autotrophic and can be outcompeted by heterotrophs. The presence of organic carbon can therefore reduce the effectiveness of the nitrification units. The ratio of carbon to nitrogen will affect which species are favoured. Especially the amount of biodegradable carbon is of interest. The ratio of biological oxygen demand (BOD) to total ammonia nitrogen (TAN) is used in this report. In order to avoid negative effects on the nitrification rate the nitrification unit is placed after the organic carbon removal unit. Nitrification: 2N H4+ + 3O2
2N O2− + 4H + + 4H2 O
2N O2− + O2
Ammonia and nitrite are toxic for aquatic animals while nitrate is much less harmful. Consequently, priorities have been on removal of ammonia and nitrite. The nitrate 7
produced is normally removed in two ways, by dilution with water exchange or by denitrification. In the denitrification, nitrate is reduced to nitrogen gas by oxidation of organic matter and is emitted to the surrounding air according to equation 2.3. Denitrification was introduced in order to increase the nitrate control and lower the water exchange rate. In cases when the denitrification does not match the nitrification the maximum allowed nitrate concentration steers the external water exchange rate in the system. The conventional semi-closed RASs have a varying external water exchange rate between 0.1-1m3 /kg feed to avoid accumulation of nitrate. This corresponds to a water renewal of 5-10% of the system volume [3, 4]. Denitrification: (2.3) N O3− → N O2− → N O → N2 O → N2 The denitrifiction occurs at anaerobic conditions by facultative bacteria. The facultative bacteria are using electron donors originated from organic or inorganic sources. In RAS and traditional wastewater treatment plants, heterotrophic denitrification is the most commonly applied method. It uses organic electron donors from a carbon source (e.g. carbohydrates, organic alcohols) that can be added externally to the system or originate from the fish feed or faeces. If the process has limited access to a biodegradable carbon source, accumulations of intermediate products, such as NO2 and N2 O, can occur. If the process has an excess of carbon, the concentration of ammonia could increase due to AOB being outcompeted by heterotropic bacteria. By reducing the concentration of nitrate, the need for water exchange will be lowered and thus decrease the water use of the process. Apart from the direct toxic effect from high nitrate concentrations on aquatic animals, there are regulations on how much nitrate that is allowed to be discharged. Since the denitrification reduces the nitrate levels and thereby the water use, these restrictions are more easily attained and increase the sustainability of the RAS . Another positive effect of denitrification is improved alkalinity. The intensive nitrification of RAS leads to a decreased alkalinity and a resulting drop in pH. Acidic conditions negatively affects the performance of the biofilter and the environment for the aquatic organisms. Alkalinity supplements, usually sodium bicarbonate, are commonly added to stabilize the alkalinity and pH. By incorporating heterotrophic denitrification the alkalinity will be increased and thus the need for alkalinity supplement will be reduced or even eliminated. There is also a risk with a low water exchange rate. When much of the same water is used in the process, accumulation of growth inhibiting substances may occur. These substances come from the fish, bacteria or the food and cannot be degraded by the water treatment processes. Examples of these substances are cortisol, a stress hormone from the fish, or metals that are brought to the process by the feed. After the denitrifying units the water is transported to aerated tanks for organic carbon removal. In these tanks organic material is consumed by bacteria and carbon dioxide is released. The tanks for denitrification and organic carbon removal are connected in series.
Ion chromatography was used in order to determine concentrations of the nitrogen containing ions in the system. However, there are disproportionate concentrations of ammonium and sodium in this system and since they have similar retention times that causes interference. In Figure 2.2a and 2.2b there is an example of a chromatogram where this can be seen. This is common when there are disproportionate concentrations of sodium and ammonium, but by using different equipment better separation of the peaks can be achieved. This was not in the scope of the project and this source of error in determining ammonium concentration could not be avoided.
(a) Chromatogram of sample before nitrification unit.
(b) Close up of ammonium peak close to sodium peak.
Figure 2.2: Example of chromatogram with interference between sodium and ammonium peaks.
As mentioned in Section 2.1, carbon is required for denitrification but undesirable in nitrification. No external carbon source apart from the fish feed is used in the 10
studied RAS. In order to determine the amount of carbon present in the system the total organic carbon (TOC) was measured. Samples were taken so that the change in concentration over the different treatment units could be determined. In order to find out how much of that carbon that could be utilized by the microorganisms biological oxygen demand (BOD) and chemical oxygen demand (COD) were analyzed. BOD is a measurement of how much oxygen is consumed by microorganisms in a sample over a specified time. BOD7, for example is the consumption over seven days which was used in this case. This can be compared to COD which is the oxygen consumption when the content of a sample is oxidized chemically.
In order to estimate the residence time in the bioreactors the hydraulic retention time (HRT) was calculated using the relation: HRT = V olume of tank/Inlet f low rate
Using the residence time along with concentrations from the flow to and out of the reactor the reaction rate can be estimated using: Reaction rate = (Cin − Cout )/HRT
As mentioned in section 1.1.1 fish excrete ammonium. However they only do this when they have been fed. When they are being fed the excretion rate increase and when the feeding stops the excretion decline over time. Approximately five hours after feeding ceased the ammonia production was undetectable in a study by Bovendeur et al.. The excretion rate of total ammonia nitrogen is estimated to be 3% of the daily feeding rate.